• by:
  • August 24th, 2021
  • Category:

First distributed learning network machines for DRAGON installed


Radiomics installed in CHU Liege and University of Florence the first machines required for the connection to the distributed learning network in the framework of DRAGON. This project, funded by the European Commission’s Horizon 2020 program Innovative Medicine Initiative (IMI) aimed at using artificial intelligence (AI) and machine learning to develop decision support systems capable of delivering a more precise coronavirus diagnosis and more accurate predictions of patient outcomes.

News

First distributed learning network machines for Covid-19 DRAGON project installed in CHU Liège and University of Florence

Radiomics installed in CHU Liege and University of Florence the first machines required for the connection to the distributed learning network in the framework of DRAGON.

Liege (BE), 24/08/2021 – Radiomics, Liege-based imaging research organization focused on AI powered healthcare through Radiomics, Deep Learning & Federated Learning, installed in CHU Liege and University of Florence the first machines required for the connection to the distributed learning network in the framework of DRAGON. This project, funded by the European Commission’s Horizon 2020 program Innovative Medicine Initiative (IMI) aimed at using artificial intelligence (AI) and machine learning to develop decision support systems capable of delivering a more precise coronavirus diagnosis and more accurate predictions of patient outcomes.

DRAGON machine CHU Liege

The DRAGON consortium, led by Radiomics, consists of 18 partner organizations, including CHU Liege and University of Florence. The project will draw on new and existing data and sample collection efforts, including CT (computed tomography) scans to carry out detailed profiling of patients. AI technology will then be used to transform this information into a precision medicine approach that will help clinicians and patients with decision making around diagnosis and treatment. Underpinning all of this will be a federated machine learning system that will allow for the use of anonymized data from a range of international sources while complying with the EU’s General Data Protection Regulation (GDPR).

As the lead partner of the consortium, Radiomics is responsible for setting up of the distributed learning network. The first milestone has been reached on July 27th, when Radiomics installed the first machine required for the connection to the network at CHU Liege. Benjamin Miraglio, Head of Technology at Radiomics, commented: “With Radiomics’ DistriM solution, datasets uploaded by a partner to its local DRAGON machine will never be transferred over the network. During the training process, AI models to be trained are downloaded from a remote server at the beginning of a training pass, and reuploaded to that server when a training pass is finished. Therefore, privacy sensitive data never leaves the data stores while available to the learning application”.

Dr. Julien Guiot, pulmonologist and head of clinics in the respiratory medicine department of CHU Liege is coordinating the project for his organization. He added: “The innovations undertaken under the DRAGON funding umbrella are particularly thrilling. It is getting clear for all that data management, and notably dedicated to clinical data, is a challenge of utmost importance. Building new paths to connect the needs of solution developers and clinical data controller, while strictly respecting GDPR rules, is necessary to benefit entirely from what technologies, and in this case AI, can offer. In order to achieve our priority goal of offering the best care possible to our patients, we, at CHU Liege, want to build trustworthy collaborations with industrial partners offering robust solutions that will add value to the clinical decisions. This project is a great illustration of my vision for the future of healthcare.

A second machine was already delivered to the partner at the University of Florence and is operational as of Friday August 20th. This represents the completion of the first phase of distributed learning installation in DRAGON which will allow for testing and optimization of the network at small scale to ensure full deployment and set up of the network between all clinical partners in DRAGON by the end of September.

Media contact

Fabrizia Candido
Account & PR Manager
[email protected]